• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.37 no.5
• /
• pp.337-344
• /
• 2024

This study presents a deep learning-based framework to predict the aerodynamic performance of low Reynolds number airfoils. The framework employs a convolutional neural network (CNN) combined with a variational autoencoder (VAE) to efficiently handle large datasets. Moreover, the signed distance function is used as the network input to represent the airfoil configuration in the image data and parameterize the CNN. A novel generative model based on projection-based manifold learning is proposed to overcome the data mining limitation of computational fluid dynamics which may incur significant computational costs. The interpolation and extrapolation accuracy of the proposed framework is evaluated using the NACA 4-digit airfoil configuration.The results show improved accuracy via data augmentation performed by the proposed generative model.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.2
• /
• pp.103-108
• /
• 2022

In this paper, we propose a new image analysis method based on regression manifold 3-D PCA. The proposed method is a new image analysis method consisting of a regression analysis algorithm with a structure designed based on an autoencoder capable of nonlinear expansion of manifold 3-D PCA and PCA for efficient dimension reduction when entering large-capacity image data. With the configuration of an autoencoder, a regression manifold 3-DPCA, which derives the best hyperplane through three-dimensional rotation of image pixel values, and a Bayesian rule structure similar to a deep learning structure, are applied. Experiments are performed to verify performance. The image is improved by utilizing the fine dust image, and accuracy performance evaluation is performed through the classification model. As a result, it can be confirmed that it is effective for deep learning performance.